+cc abarth, epoger, reed On 2011/10/06 16:46:37, Nicolas Sylvain wrote: > 10% slowdown on a ...
9 years, 2 months ago
(2011-10-06 17:37:34 UTC)
#3
+cc abarth, epoger, reed
On 2011/10/06 16:46:37, Nicolas Sylvain wrote:
> 10% slowdown on a webkit roll? that seems like a big deal. Are there people
> fixing it?
I have cc'd some people from the Webkit/Skia teams. Can one of you inform us
that at least the major regressions (O(10%)) are being examined and are going to
be fixed in the next day or two?
I've been lenient in letting people try to fix their issues in follow-up
changes. But the fixes for big performance regressions were not provided
rapidly. So I'm now going back to pressing for reverts to these changes once we
identify the causing CL.
Adam probably has a good opinion about what to do here. Adam, WDYT about these
regressions?
Nicolas Sylvain
On Thu, Oct 6, 2011 at 10:37 AM, <cmp@chromium.org> wrote: > +cc abarth, epoger, reed ...
9 years, 2 months ago
(2011-10-06 17:59:49 UTC)
#4
On Thu, Oct 6, 2011 at 10:37 AM, <cmp@chromium.org> wrote:
> +cc abarth, epoger, reed
>
>
> On 2011/10/06 16:46:37, Nicolas Sylvain wrote:
>
>> 10% slowdown on a webkit roll? that seems like a big deal. Are there
>> people
>> fixing it?
>>
>
> I have cc'd some people from the Webkit/Skia teams. Can one of you inform
> us
> that at least the major regressions (O(10%)) are being examined and are
> going to
> be fixed in the next day or two?
>
> I've been lenient in letting people try to fix their issues in follow-up
> changes. But the fixes for big performance regressions were not provided
> rapidly. So I'm now going back to pressing for reverts to these changes
> once we
> identify the causing CL.
>
> Adam probably has a good opinion about what to do here. Adam, WDYT about
> these
> regressions?
>
For what I can see, the regression is in 102911, which is a webkit roll
(r96065:96091). There are 26 changes in that range, but AFAIK only 9 that
applies to us:
r96066 | jamesr@google.com
r96068 | ggaren@apple.com
r96069 | simon.fraser@apple.com
r96070 | simon.fraser@apple.com
r96073 | haraken@chromium.org
r96078 | rakesh.kn@motorola.com
r96082 | bmf834@motorola.com
r96084 | rniwa@webkit.org
r96087 | carlosgc@webkit.org
>
http://codereview.chromium.**org/8166014/<http://codereview.chromium.org/8166...
>
abarth-chromium
> Adam probably has a good opinion about what to do here. Adam, WDYT about ...
9 years, 2 months ago
(2011-10-06 20:01:46 UTC)
#5
> Adam probably has a good opinion about what to do here. Adam, WDYT about
these
> regressions?
I don't have any context this for this particular change, but we have been too
lenient on performance regressions in the past couple months. I would put
pressure on folks to not regress performance.
abarth-chromium
Ah, now I know which regression you mean. We're tracking this as https://bugs.webkit.org/show_bug.cgi?id=69238. This is ...
9 years, 2 months ago
(2011-10-06 20:04:26 UTC)
#6
Ah, now I know which regression you mean. We're tracking this as
https://bugs.webkit.org/show_bug.cgi?id=69238. This is the one I've been
emailing you about tracking down. We're not sure which of those changes caused
it.
Issue 8166014: Update expectations.
(Closed)
Created 9 years, 2 months ago by cmp
Modified 9 years, 2 months ago
Reviewers: Nicolas Sylvain, abarth-chromium
Base URL: svn://svn.chromium.org/chrome/trunk/src
Comments: 0